Recent Returns Are a Lagging Indicator. The Liquidity Cycle Leads.

Recent Returns Are a Lagging Indicator. The Liquidity Cycle Leads.

The Cleveland Fed went and did something rare: it asked thousands of American households, repeatedly, what they actually do with cryptocurrency. The finding now making the rounds is that investors are swayed by Bitcoin's recent returns — show them a run of strong past performance and they want more crypto; hand them a losing tape and they're out. On its face, that is one more data point for the crowd that insists crypto is a momentum sideshow driven by recency bias and animal spirits.

Hold that thought, because the same instinct is running through the market right now. Bitcoin is up roughly 20% in a month, and fear-and-greed gauges have flipped to, well, greed. Yet the conclusion most readers will draw from this study — that crypto is froth — is, in my view, the exact opposite of the useful takeaway. What the study actually documents is which variable investors anchor to. And it is the lagging one.

What the survey actually found

The paper — "Do You Even Crypto, Bro? Cryptocurrencies in Household Finance," out of Cleveland and also circulated as an NBER working paper — describes an owner base you would recognise from any macro survey of this generation: young, male, more libertarian than the average, expecting far higher returns than non-holders think are coming, and — tellingly — perceiving the asset class as relatively safe. Then comes the experiment: when the researchers fed people information about historical crypto returns, desired holdings went up, and actual purchases followed.

The finding that does not make the headline round is the more structural one. Fluctuations in Bitcoin's price directly change how much crypto holders spend on durable goods — the washing machines, cars, and fridges of the world. The wealth effect now runs from the Bitcoin ledger straight into consumption. Whatever you think of the asset, that is a real-economy transmission channel. Crypto, in the data, is already macro.

The interesting part is what the survey cannot see

The temptation is to read all of this as confirmation that crypto is a speculative toy for extrapolating gamblers. That conflates the survey's subject — household expectations — with the thing that actually drives the asset. Think about the information problem first. Bitcoin pays no coupon, no dividend, no earnings stream. There is no discounted cash flow to anchor an expectation to. For a household, the only fundamental that exists is the price history. Of course they extrapolate it. The study is describing the only tool a retail investor has, not proving that the tool is a delusion.

What no household survey can measure is the independent variable. The driver of this asset class is not recent returns. It is the aggregate of central-bank balance sheets, money supply and credit creation — the global liquidity cycle. That is the framework that has held up best here: Bitcoin's implied ISM tracks the actual ISM, and the trend in Fed Net Liquidity — the Fed's balance sheet minus the Treasury General Account and the reverse repo facility, the practitioner's measure of cash actually available to risk markets — has described risk assets in general and digital assets in particular. The price is the visible readout of an invisible driver. The crowd watches the readout. What matters is what the driver is doing.

The market is running the Fed's experiment in real time

And right now we have a live test. Look at the tape first:

Period	Bitcoin, Aug 23 2026
5 days	+19.5%
20 days	+21.8%
60 days	+26.7%
1 year	−12.0%
Year-to-date	−6.6%

So the story is not "crypto raging." The story is "crypto down 40% from its high, then ripped more than 20% in a month off the lows" — roughly $77,000 today against a 52-week low near $57,800 and a high near $125,500. The altcoin-season gauge sits at 31 with Bitcoin dominance near 60%, which is the signature of liquidity rotating into the deepest, most liquid asset first, not an everything-frothy blowoff. Turnover has exploded in the spot market: gross inflows into the biggest BTC pair spiked above $1.3 billion, then $1.6 billion, on consecutive days this month, even as net flows stayed mixed — the heavy churn of the momentum crowd re-entering. That is precisely the "recent returns" the Cleveland Fed's households are wired to extrapolate. The study's subjects are showing up in real time.

But here is what the survey cannot tell you, and what will decide whether the chasers get paid. The liquidity cycle that dominates this whole trade is inflecting in the constructive direction. The Fed, having grown its balance sheet from roughly $800 billion in 2005 to about $6.5 trillion — from 6% to 21% of GDP — judged by December 2025 that reserves had fallen to efficient, ample levels, and the multi-year runoff has been winding down since.

The funds rate has sat on hold in a 3.50%–3.75% range through the summer. US M2 set a fresh record of $23.2 trillion in June, and the combined money supply of the four major central banks was running 8.1% higher year on year in June. Global liquidity is expanding. That is the leading indicator.

Now look at the geometry. The recency crowd is chasing a 20% month into an expanding liquidity cycle — the best possible conditions for a momentum chaser — and that is exactly why their frame is fragile. Their horizon is two days long. The liquidity cycle's horizon is quarters. The moment the tape stalls for a week, the same anchor that pushed them in will push them straight back out.

That is why sentiment extremes are the tradable signal, not the tape

Which brings me to why this study, properly read, reinforces the contrarian framework rather than indicting crypto. Recency bias is a lagging indicator by construction. It tells you what the crowd did with the last move; it says nothing about what the next move does to them. Sentiment extremes only become a useful contrarian input when a leading indicator confirms them.

The cleanest demonstration was Q4 2022, when literally everyone was bearish — because the recent returns were terrifying. The recency logic at that moment said hide, stand down, do not touch risk. The global liquidity cycle bottomed in October 2022, a month before the rally that followed. The crowd anchored to the lagging variable was wrong at precisely the moment that mattered most — and right for most of the rest of the time, which is exactly why it stayed anchored.

The honest counterpoint is that this study is the cautionary tale, not the vindication. A household base that extrapolates a 20% month is the same fuel that evaporates the moment M2 rolls over. True. Which is why you cannot trade recency alone. The extrapolative base only gets paid when the lead indicator backs it up; if global money growth stalls and the Fed's balance sheet resumes shrinking, the chase runs just as fast in the other direction.

So what does the macro lens say to watch? The data would break this read if global M2 rolls over, if the Fed resumes balance-sheet shrinkage, or if the reverse repo and Treasury account start draining liquidity back out of the system. I flag one gap plainly: household positioning is measured with a lag — the survey publishes on its own schedule, and the lab experiment moves desired holdings, not necessarily the billions that actually traded this month. On the information I have, though, the leading indicator and the positioning are not in conflict the way they were at prior tops.

Do not let this study sell you the story that crypto is steered by recent returns — that is the crowd's information set, not the mechanism. The mechanism is liquidity, and liquidity is currently doing the work that makes a recency-chasing crowd look smart, for as long as it lasts. Crypto is macro and macro is crypto: the crowd watches the return, and the return always follows the liquidity. Get the liquidity right, and you get the crypto right.